DIFFERENTIAL EQUATIONS
Structure
14.1 Introduction
Objectives
14.2 The Complete Integral
Compatible Systems of First Order Equations
Charpit's Method
Standard Forms
Jacobi's Method
14.3 Cauchy Problem
14.4 Summary
14.5 SolutionslAnswers
14.1 INTRODUCTION
In Unit 13, we considered only linear PDEs of the first order. From the discussion in the
last unit you must have observed that the construction of an integral of linear PDEs of
first order is a multistage process and in this respect it differs from the usual construction of
an integral of an ordinary differential equation. The problem of finding the integral of a
non-linear PDE of the first order is more involved than that for the corresponding
quasi-linear equation although there are some striking similarities. We shall discuss in this
unit the general method of finding the complete integral of non-linear PDEs of the first ordkr.
--
The method of finding the complete integral of non-linear PDEs of the first order is partly
due to the Italian mathematician Lagrange (1736-1813). Later on, it was the French
mathematician Charpit who perfected it and presented it in a memoir in 1784 to Paris
Academy of Sciences. The method is based on the consideration of compatible system of
first order equations. We have, therefore, started the unit by defining the compatible system
of equations and obtained the conditions for systems to be compatible.
The methods due to Charpit and Jacobi, as you shall see later in this unit, aim at constructing
the integral containing a number of arbitrary constants. The results obtained by these
methods do not indicate any particular suggestion of Cauchy's theorem and do not help in
finding a solution to initial-data problem. Consequently, we shall state in Sec. 14.3 the
Cauchy problem, which is based on Charpit's method and which gives the solution of
non-linear first-order PDEs satisfying initial data.
Objectives
.After
-
studying this unit, you should be able to
a define compatible systems of first order equations and find the condition for two
systems of first order non-linear PDEs to be compatible;
a' use Charpit's method for finding the complete integral of a non- linear PDE of first
order;
.-
a identify special types of equations for which short methods can be used to determine
their integrals;
a use Jacobi's method for solving first order non-linear PDEs;
.-
a state the role of the complete integral in determining the integral surfaces of the
non-linear PDE for the Cauchy problem.
I Flrst Order Partial Dlfferentiel
Equations 14.2 THE COMPLETE INTEGRAL
has been developed by Charpit. This method consists in finding another PDE which is
compatible with the given Eqn. (I) and which involves an arbitrary parameter.
Before taking up the Charpit's method, we shall define when a system of first order PDEs
are compatible and what are the conditions satisfied by such systems.
g(x,y,z,p.q) = 0.
we say that Eqs. (1) and (2) are compatible.
We can solve Eqns. (1) and (2) to obtain explicit expressions for p and q in the form
The condition that the pair of Eqns. (1) and (2) should be compatible, then, reduces to the
condition that the equation
Ybu know that Eqn. (5) is a Haniaf differential equation and the necessary and sufficient
condition for its integrability is (ref. Unit 12, Theorem 1)
where X = (@.V,-1).
Now, X . curl X = ($ i + v j - k) . [ i (-vZ) + +k (vx - $y)l
Thus Eqs. (1) and (2) are compatible if Eq.(6) is satisfied where 4 and yf are given by Eq.(3).
In order to obtain the compatibility condition in terms off and g, we substitute in Qs.(l)
and (2) the values of p and q from Eq.(3) and then differentiate the resulting Eqs.(l) and (2)
with respect to x and y, and obtain 1
Non-linear Partial Differential
Equations
Multiplying Eqns. (7) and (8) by gpand fp respectively, and subtracting the resulting
equations, we get
. where we have used relation (3) and replaced $I by p in the second term.
I Similarly, multiplying Eqs.(9) and (10) by g,-,and fq respectively, and subtracting the
resulting equations, we obtain
f Adding Eqs.(l 1)'and (12) and using relation (6). we obtain the compatibility condition for
Eqs.(l) and (2) as
The expression on the left hand side of Eqn. (13) is denoted by [f,g]. Thus, we have
We can use Eq.(14), which is a first order PDE,for finding the equation g = 0, compatible
with the given equation f = 0. Once g is known we can find p and q and then integrate Eq.(5)
to obtain one parameter family of solutions in the form
where b is an arbitrary constant. The solution given by relation (15) shall satisfy both
Eqs.(l) and (2). Thus, compatible equations have a one-parameterfamily of common
solutions.
Let us now consider a few examples.
dz=pdx+qdy,
we obtain
f(x,y,z,p,q) = 0
that is homogeneous in x, y and z.
E 3) Show that the equations (y-z)p + (z-x)q = x-y and z-px-qy = 0 are
compatible.
E 4) Verify that the equations xp - yq =x and xZp + q = xz are compatible and find
their one-parameter family of common solutions.
We shall now discuss Charpit's method of finding the complete integral of Eq.(l).
14.2.2 Charpit's Method on-linear Partial Differential
Equations
As mentioned earlier, in the Charpit's method of finding the complete integral of non-linear
PDE of the form (I), namely,
f(x,y,z,p,q) = 0,
we introduce another PDE of the f i s t order of the type
should be integrable.
Once we are able to find such a function F, we can integrate Eq.(24) and obtain the solution
in the form
You know that Eqns. (1) and (22) are compatible only if Eqns. of the type (4) and (14) are
satisfied, i.e., if
and
Eqn. (26) is a first order linear PDE for the function F which y/e have considered to be a
function of five variables x,y,z,p and q. Following Lagrange's method of solving first order
linear PDE, the auxiliary equations for Eqn. (26) are
Note that we can write down Eqns.(27) at once from the given Eqn. (1). Eqns.(27) are the
characteristic differential equations or Charpit's equations related to the given PDE
(1).
If we can find a single solution of the system of Eqns.(27)'involving p or q or both in the
form of Eqn. (22), viz.,
then the problem reduces to solving Eqns.(l) and (22) for p and q and then integrating
Eqn. by the methods of solving Pfaffian differential equations.
Remark :It is not necessary to use all of the Charpit's Eq.(27) for any given problem. What
you have to be careful about is that p or q must occur in the solution obtained.
We now take up an example to illustrate this method.
First Order Pnrtial Differential Example 3 : Find the complete integral of
Equations
z2 = pqxy
by Charpit's method.
t pqxy - z2 = 0.
Solution :~ ef =
In this case, the auxiliary Eqns.(27) yield
--
dx d p = dq
qxy -pxy - 2pqxy - 2zp-pqy 2zq-pqx
dz dz pdx+qdy+ xdp+ ydq
=$ ---- (using given equation)
~ W X-Y2z2 - ~Z(PX+~Y)
where a(c-' + c ) = 1.
Example 4 : Using Charpit's method find the complete integral of the equation
2 2
(P +q ) Y= q z
Solution : The given equation can be expressed as
2 2
f = ( p +q ) y - q z = O ,
The auxiliary equations in this case are
Substituting these values of p and q in dz = pdx + qdy and taking only positive sign, we get
Non-linear Partial DIfPerential
Equations
I Integrating, we get
i
I which is the required complete integral.
And now an exercise for you.
E 5) Using Charpit's method, find the complete integrals of the following equations :
After solving E 5), you must have observed that the general method given by Charpit for
finding the complete integral of a non-linear first oider PDE is usually quite lengthy and
involved. However, there are some special types of first order non-linear PDEs whose
complete integraIs can be easily obtained by Charpit's method. These special types of
non-linear PDEs of first order are called standard forms of Eq.(l). We now take up the
methods of integrating these standard foms.
One obvious solution of the above system of equations is obtained by considering the first
and the fourth fraction. This solution gives us the compatible equation
p = constant = a, say.
Once the value of p is known to us, we can obtain thecorresponding value of q from Eqn. (30)
in the form
q = Q(a), a constant.
Therefore, the equation
First Order Partial Differential dz = pdx + qdy
Equations
reduces to
dz = adx + Q(a)dy
We may remark here that, in the case under consideration we have chosen equation dp = 0
from the system of auxiliary equations to obtain the second compatible equation. in some
situations it is possible that the computations involved are reduced if we take dq = 0, leading
to q = a constant.
Let us take up some examples.
Solution : For the given equation, we consider only thc following Charpit's equations
3p = constant = a, say.
dz = pdx + qdy,
we obtain
We next take up an example to illustrate how a given equation can be easily solved if we
first reduce it to the form (30).
Then
and Non-linear Partial Differential
Equations
Substituting for p and q from relations (32) and (33) in the given equation, it reduces to
This equation is of Type 1 and we can write down its complete integral directly by using
Eqn. (31) in the form
=a G + aG+
b (substituting back the values of U and Y in terms of r and
Y.)
You may now try the following exercise.
-
We next take up those equations which do not involve the independent variables x, v
explicitly.
f(2,p.q) = 0
Charpit's Eqns.(27) in this case assume the form
P = aq, . . . . (35)
where a is a constant.
f(z,aq,q) = 0
First Order Partial Differential dz = pdx + qdy,
Equations
we get
dz=aqdx+qdy
= (adx + dy)q
zpq-p-9'0
Solution : The given equation is
It is a non-linear PDE of the first order not involving the iddependent variables. In this case
we consider, as above, the Charpit's equations
On integration, we get
Now, i f q = O , p = O
dz = 0
=$ z = constant.
which is obviously not the complete integral of the given equation.
a+ l
Further. for q = -- - ,p = P+I. the relation
az z
dz = pdx + qdy
yields
which on integration gives the complete integral of the given equation in the form Nun-linear Partial l)iffere~~tial
Equations
We next take up the equations in which z does not occur explicitly and which are separable.
In other words, we consider the equations of the form which can be written as
f(x,p) = g(y 4)
Type 111 :Separable Equations
You may note that Eqn. (41) is an ODE in x and p. We can solve it for p by writing it in the
form
Any combination of + signs can be taken. We take only the +ve signs here and then
yields
You may recall that in Unit 1, Block 1, we defined the ODE of the type
Y = XP + f(p)
as Clairaut's Equation.
In the case of a first order PDE, we can write the Clairaut's equation in the form
We now give you the method of solving Clairaut's equations by using the Charpit's
equations.
px + qy + f(p,q) - z = 0
Non-linear Partial Differential
For Eqn. (44), the Charpit's Eqs.(27) take the form Equations
_dx-- = - dz --!-!E!=@
-
x+fp y+fq px+qy+pfp+qfq 0 0
p=aandq=b
z = ax + by + f(a,b),
which is the complete integral of 'Eqn. (44).
From 1st and 4th fractions and 1st and 5th fractions, we get'
dp=O,dq=O
Integrating, we get
p = a and q = b,
We have already mentioned in Sec. 14.1 that Charpit's method cannot be generalized for any
number of independent variables directly. Consequently, we take up, in the next subsection a
method, due to Jacobi, for solving PDE of the form (1) which can be generalized for any
number of variables.
First Order Partial Diflerential 14.2.4 Jacobi's Method
Equations
Jacobi's method of integrating the PDE of the form (I), viz.,
u(x,y,z) = 0
then, we have
where u denotes'-
au ( i=1,2,3, with xl = x,x2 = y and x3 = z )
axi
Substituting from Eqns.(46) into Eqn.(l), we obtain a partial differential equation of the type
You may note here that the advantage of having Eqn.(47) is that in it the dependent variable u
does not appear explicitly. a
'-1n Charpit's method, we tried to obtain one PDE of the first order compatible with the given
equation. In Jacobi's method we look for the two PDEs pf the first order of the type
and .
1
You know that the complete integral of Eqn. (1) requires only two arbitrary constants. The
solution obtained by Eqn. (50) will therefore give us different complete integrals of the given
Ean. (1). Thus the complete integral of Eqn.(l) is obtained by making different choices of the
third constant.
We now aim at determining two Eqns. (48) and (49) which are compatible with Eqn. (47).
.
The compatibility conditions(l3), in the case of Eqns. (47) and (48) take the form
Eqn. (5 1 ) is a linear PDE of the first order and using Lagrange's method, its auxiliary
equations are
Similarly, for Eqs.(47) and (49), we have the compatibility condition as Non-linear Partlal Differential
Equations
d(F.h) + a(F.h) + 2F.h) -
~(x.u a(y 4,) a(z.u3)
Once we have written down Eqn. (1) in the form of Eqn. (47), we can at once write down the
system of Eqn.(52). The auxiliary functions g and h, given by Eqns.(48) and (4% can then
be obtained as the two solutions of the linear PDEs (52).
As mentioned earlier, the chief merit of Jacobi's method is that it can be generalized for any
number of variables. If we have to solve an equation of the type
fl(xI,x2,.. . . .,Xo.U,,U2,. . . . , u n ) = O . . . . . (53)
au
where u denotes -(i = 1,2,...,n), then we can find (n-1) auxiliary functions f2,f3, .. ..
axi
f,-, ,f, involving (n-1) arbitrary constants, from the equations
where, (fl)
I'
Jfl
denotes -, i = 1,2. . ,n.
aui .
.
Once we have solved Eqns. (53) and (54) for ul,u2,. .. . .,un, we determine u by integrating
the Pfaffian equation
n
du = u, dxi . . . . (55)
I= 1
The solution so obtained willcontain n arbitrary constants. This type of direct generalization
is not possible in the case of Charpit's method. But, the main difficulty in Jacobi's method is
that here we are required to find two compatible systems, which ~t times is quite
cumbersome.
z ~ + z u ~ - u ~ ~ - u ~ ~ = O
Solution : This equation is already in the form of Eq.(47). The auxiliary Eqns.(52) reduce
to
--
dx - d y - -d z - du1 - du2 - du3
-2ul -2u2 z 0 0 -2z-u3
.is\ Therefore, from the first and fourth and first and fifth ratios, we get
ul = a, u2 = b, a and b being constants.
With the above forms of ul and u2, we obtain from the given equation
a2 + b2 -z2
u3=-
Z
Now,
On integrating the above equation, we arrive at the following solution of the given equation
1
u=ax+by+(a2+b2)l n ~ z ~ - = z ~ + c .
First Order PartialDifferentlel where c is the constant of integration.
Equetlons
Example 11: 'Find a complete integral of the equation
by Jacobi's method.
u1 u2
Solution : Writing p = - - ,q = - -,
u3 u3
the given equation is transformed to
we obtain
Ybu have seen that Charpit'r; and Jacobi's methods of solving non-linear PDEs of the first
order aim at constructing co~npleteintegrals containing a number of arbitrary constants.
However, these results do not indicate a solution of initial data problern. Consequently, in
the next section. we shall state the Cauchy problem which is based on Charpit's method for
solving non-linear PDEs of the first order. satisfying initial data.
14.3 CAUCHY PROBLEM 'C
consists in finding the integral surfaces Si of Eqn. (I), which pass through the initial data
curve To given by
Note that due to non-linearity of Eqn. (1). there can be more than one integral surface
corresponding to the data (56).
The initial values po(s), qo(s) defined along To must satisfy the equations
f(%(s), y0(s). ZQ(S).p0(s). qo(s)) = 0 ....(58)
and
Note that Eqns.(58) and (59) may have several pairsbf solution pq(s),&(s)depnding on the
non-linearity of Eqn. (58).
Also, from Eqns.(27), you know that the.slope of the chaiacteristic curves of the given PDE
is
Since the initial data curve To should be nowhere tangential to the characteristics of the
PDE, therefore, the integral surfaces Si corresponding to the initial data (xo(s), yo(s), q,(s).
poi(s), Q~(s))must satis@the condition.
Step I : Write the auxiliary Eqn.(27) for the PDE (1). namely,
Step I1 : Choose suitable fractions from these equations to find the compatible family of
Eq.(22), viz.,
F(x.y,z,p,q.a) = 0
The condition
First Order Partial Differential will be automatically satisfied on some domain D.
Equations
*
Step 111: Solve Eqn. (1) and (22) for p and q to get Eqns.(23), namely,
P = @(x,y.z,a),q = v(x,y,z,a)
dz = $(x,y,z,a& + \y(x.y,z,a)dy
and integrate to ob%n the complete integral of Eqn. (1) in the form (25), namely,
G(x,y,z,a,b) = 0,
where a and b are arbitrary constants.
Step V : Substitute from the initial data curve To in Eqn. (25) to get
Eliminating s between Eqns.(62) and (63), we get a relation between the constant a and b in
the form
Step VI :Obtain the envelopes of Eqn.(66) by eliminating the parameter 'a' between
Eqn.(66) and
These envelopes are the integral surfaces Si, i = 1,2, . . . ,n, all of which pass through the
initial curve T o but corresponding to different ( poi(s),qoi(s)).
We iIlustrate steps I-VI above by the following examples.
Example 12 : Find the complete integral of the equation
(p2+q2)x= pz
and the integral corresponding to the initial data,
t T 0 : x , = O , y o = s2,z0=2s.
Solution :Step I : The auxiliary Eqn.(27) for the given equation are in the form
Step I1 : From the last two fractions, we obtain the equation which is compatible with the Non-linear Partial Differential
Equations
given PDE in the form
p2+q2 = a2, a being a constant.
2
p = -a, qx= f a &TT
z-ax
Z Z
zdz - a2x dx -
-f ady
GP
On integrating the above equation, we obtain the complete integral in the form
z2 = a2x2+ ( a ~ + b ) ~ , . . . . (69)
where b is the constant of integration
Step V : Substituting the initial data in Eqn. (69). we get
4s' = (a~'+b)~
Differentiating Eqn. (70) with respect to s, we find
Eliminating s from Qns. (70) and (71). we get a relation between a and b of the form
a b = 1.
1
Substituting b = - in Eqn. (69). we get the one-parameter subsystem of the complete integral,
a
i.e.,
Step VI :We now obtain the envelope of -Eqn.- (72) by eliminating a between Eqn. (72) and its
partial derivative with respect to a, i.e.,
The envelope is
( 2 y - ~ ~= )4~(x2+y2)
$1 and is the required surface. We can rewrite this equation in the form
Since fi
x + y 2 y, we discard the minus sign for the real solution.
Thus,
z(1-q2) - 2(px+qy) = 0
which passes though the line x = 1, y = hz + k has the equation
(y - kx) = z i(l+h2)x - l]'12.
Solution :Step I: The auxiliary Eqn.(27), for the given equation are in the form
Step IV : Substituting from Eqns. (75) and (76) in dz = pdx + qdy, we get
zdz-ady --
- dx
~o:~=l,y=h~+k,z=~, . . . .(78)
where s is the defining parameter,
Substituting from relation (78) ,intoEqn. (77). we obtain
s2 - 2a(hs+k) - a2 - 2b = 0 ... . (79)
Diffefentiating Eqn. (79) with respect to s, we find
s-ah=O
Eliminating s'from Eqns. (79) and (80). we find
2b = -a2 (l+h2) - 2ak . . . . (81)
Substituting the value of b from relation (8 1) into Eqn. (77). we get
z2 = a2+ 2ay-x [a2(l+h2)+2ak] . .. . (82)
Step VI : The envelope of Eqn. (82) is
E 1 1 ) Find the complete integral of the equation p2x + qy - z = 0 and derive the equation
of the integral surface of which the line y = I, x+z = 0 is a generator.
E 12) The normal to a given surface at a variable point P meets the sphere x2+y2+z2=
I in the points A and B. If AB is bisected by the plane z = 0, show that the
surface satisfies the PDE
z(p2+q2)+ px + qy = 0
We now end this unit by giving a summary of what we have covered in it.
14.4 SUMMARY
t
2. The Charpit's method of solving PDE of the form ( 1 ) consists of the follo ing steps.
b)
F(x,y,z,p,q,a) = 0.
to solve f = 0 and F = 0 for p and q in terms of x,y,z.
I~
involving p or q or both which contains an arbitrary constant and is of he form
~
~~
C) substitute the above values of p and q in'the relation
, dz = pdx + qdy
and integrate it to obtain the complete integraI.
h.i
\
3. There are fouf speciaI forms of Eqn. ( I ) called standard forms to which C arpit's
method can be easily applied. These forms are
a) Equation involving p and q only, for which the auxiliary equation reduc to
b) Equations not involving the independent variables, for which the auxilia
reduce to a equations
First Order Partial Differential
Equations
where a is a constant.
d) Clairaut's form of first order PDE, namely, z = xp+yq+f(p,q)for which the auxiliary
equations are
where ui =
au
axi, i = 1.2.3 with xl = x, x2 = y, x3 = z
F ( x , ~ , z , u ~ , u ~=. 0
u ~(see
) Eqn. (47)).
In order to solve Eqn.(l), we write down auxiliary equations for F(x,y,z,ul,u2,uj)= 0, and
then
c) Integrating du = uldx + u2 dy + u3 dz
and ldy making different choices of the third constant of integration in the solution
obtaibed, we determine different complete integrals of Eqn. (1).
14.5 SOLUTIONSIANSWERS
E 1) From Eq.(14), the compatibility condition for two PDEs of first order is
\
[f, g] = x(2xp - z) - y.(O) + [xp - yql(- x) - [p - 1 + p.0]x2 - [- q + .0]. 1
=~ X x2p
X ~ . ~ - - Z + xyq - px 2+x2+q
= 0 (using f = 0)
3 ( 1+xy)
(xdy + ydx) = 0
d(z-x) - (z-X) d( 1+xy) = 0
Integrating, we get
z-x
--- -b
l+xy
3 z = x + b(l+xy), for b being an arbitrary constant, is a one-parameter family of
common solutions of the given PDEs.
E 5) a) The given PDE is
2
p x+q y=z
2
. . . . (83)
The auxiliary equations are
--
dx - & -
- dz -
-dp=dq
2px 2qy 2(p2x + q2y) p-p2 q-q 2
where a is a constant.
On solving Eqns. (83) and (84) for p and q, we get
we get
Integrating, we get
2*+*=0
Y P
Integrating, we get
2
Y P=a . . . . (86)
Solving Eqns. (85) and (86) for p and q, we get
adx ax aL
3 ydz+zdy=--~dy+~dy
Y Y 2~
Integrating, we get
(a-x1~+2~~+2z
dz = (a-x) dx - dy
Y,
3y
2
dz = y2(a-x)dx - (a-x) 2 ydy - 2y3dy - 2z ydy
y2dz + 2zydy + (a-x) 2 ydy - y2 (a-x)dx + 2y 3dy = 0
or
Integrating, we-get
2 2
y [2z + (a-x) + y2 ] = b,
which is the complete integral of the given PDE.
d) z2 = 2 (a2+1)x2+ 2ay + b
e) z = ax + bey (y+a)"
E 6) a) The given PDE is
* p = constant = a, say
From the above equation and the given equation,we get
a+q-aq = 0
c) z = a x + e $ l + b
This'equation is now of the Type I and we can write its complete integral
directly, using Eq.(31), in the form
z = ax + (-a2)y +b
On integration, we get
. . . . (91)
From Eqns. (90) and (91), we get
C) (1+a2) (c2-z2) = ( a x + ~ + b ) ~
d) 4(a-I)z = ( a ~ + y + b ) ~
e) The given PDE is
q2y2 = z(z-xp)
Let V = Inx and V = Iny
First Order Partial Differential substituting for p and q in the given PDE, we get
Equations
It is now a first order PDE in which independent variables are not explicitly
present and its complete integral becomes
dz=pdx+qdy, we get
(P+@(z-xP-y 4) = 1
It can be written in the form
Here 1st and 2nd fractions and I st and 3rd fractions yield
Integrating, we get
b) z = a r + b y + d u a 2 + g b 2 + y
C) z = ax +by + 3(ab)'I3
E 10) a) ' The given PDE is
2
z + 2u3- (u1+uZ) = 0
The auxiliary Eqs.(52) reduce to
From the 1st and 4th and 1st and 5th ratios, we get
u, = a and u2 = b,
where a and b.are arbitrary constants.
With the above values of ul and u2, the given PDE gives
a+b) -z
du=adx+bdy+Gdr
On integration, we obtain
I
First Order Partial DiNerential where c is another constant of integration.
Equations
Now, using u = c, we get the complete integral as
p = (z + q ~ ) ~
Writing p = - -
1
.q u2
= - - ,the given PDE reduces to
2 2 2 2
3 z u3+y u ; ! - ~ ~ z u ~ u ~ + u ~ u ~ = O . . . . (92.
I
Y u2
- (2bz-a) f
2z2
Replacing the values of ul, u2, u3 from above in the relation
du = U I dx + ~2 dy + u3dz,
we get
On integration, we get
u=ax+bInI dz+c
a x + b l n I y ~ + b I n I z I + a- +
22 5 +--
a“.; dz = 0
Non-linear Partial Differential
C) The complete integral.of the given equation is
Equations
Step 111 : Solving for p from Eqn. (93) and the given equation, e find
r o : x = s , y = l,z=-s.
Substituting in Eqn. (94). we find
(a+b+2s)' = 4bs
Step VI : The envelope of Eqn. (96) is obtained by eliminating between Eqn. (96) and
which yields
xy = z(y-2)
and this is the required surface.
E 12) Let z = f(x,y) be the surface. The equations to the normal
at ( x l . ~ l , ~are
l)
where (pl,ql,-1) are the direction ratios of the normal. The line 8) meets the
cnhprp Thprpfnrp
First Order Partial Differential
Equations
If the roots of this equation which c,orrespond to the points A and B are rA and rB,
then
Now,
ZA = z1 -TA and zg = z1 -rB
and
From'Eqns.(99)
- . and (100) and dropping the suffix l, we get
z(p2+q2) + px + qy = 0; which is the PDE satisfied by the surface.
E 13) a) It is readily shown that the initial values for the given problem are
~~=v,y~=o,z~=o,p~=o,q~=2v,~=o
Step I : The auxiliary Eqn. (27) in this case are
Step 111: Solving for q, by using the above value of p in'the given PDE, we get
On integration. we get
2 ( a2+[2z-2ay-(x-a)2] )IR=2~+2b
Non-linear Partial Differential
Equations
which is the complete integral.
a' - (v-a) 2 = b2
-2(v-a) = 0
a v = a and hence a2 = b2
-b=fa
a2 + 2z - 2ay - (x-a) 2 = ( y h )2
Step VI : The envelope of the above equation is obtained on el nating a from the
above equation and the relation
--
2a - 2y + 2(x-a) = f i(y
f a)
a = 2y-x
1 2
. z=-(4xy-3y-),
2